Reproducing results in any science is quite challenging. A recent 2016 survey by Nature (http://www.nature.com/news/reality-check-on-reproducibility-1.19961) has shown 2/3 of researchers are concerned about science reproducibility. In the field of metabolomics, for results to become reproducible, descriptions of an investigation in a manuscript are insufficient. To surpass this, and increase the chance of result reproducibility, standard frameworks for data sharing and sharing of experimental data are invaluable. In this presentation, developments in data standards initiatives in metabolomics, including nmrML for NMR raw data (COSMOS initiative) mzTab developments for metabolite identification and qcML for data quality (both joint efforts by MSI and HUPO-PSI) will be discussed. It will also be shown how emerging metabolomics data sharing platforms can promote open, accessible data sharing standards. Finally, our own experiences, as well as community efforts in creating metabolomics data analysis workflows, particularly in Galaxy and KNIME environments which can capture study-specific experimental parameters will be presented. Such workflows would ideally run on a dedicated e-infrastructure platform, such as the ones currently under development by the PhenoMeNal consortium (http://phenomenal-h2020.eu/home/). Such efforts coupled with wider community involvement can pave the way for a greater reproducibility of the results in data analysis, data integration and reuse of data in metabolomics.
Previous Article in event
Next Article in event Next Article in session
Towards standard, accessible and reproducible Metabolomics.
Published: 01 November 2016 by MDPI in The 1st International Electronic Conference on Metabolomics session Analytic Techniques in Metabolomics
Keywords: standards, data sharing, reproducibility